(...) For decades, fire investigators claimed that, through experience and professional insight, they could tell a "pour pattern" from an accidental burn pattern. The process is termed "floor pattern analysis." This expert 'knowledge' was permitted as testimony in court, and often formed the basis for both criminal and civil claims.
This is the sort of evidence that killed Todd Willingham.
As you might have guessed, there's really no way to determine based on sight whether a pattern comes from an accelerant or the naturally occurring heat radiating in a fully developed fire. This was demonstrated through a series of test burns conducted in 1997 by fire scientist James Shanley, which showed that the radiating heat in a fully developed fire will not only create its own floor patterns, but can also obliterate any pour patterns that might have been present at an earlier stage of the fire.
The only way to accurately and conclusively determine if there was an accelerant present in a fire is, quite simply, to test for it in the debris analysis. This is surprisingly simple to do, with proper equipment (a gas chromatograph-mass spectrometer). Any testimony that an accelerant was used that isn't accompanied by either a.) a GC-Mass Spec result or b.) a signed and sworn confession ought to be treated as speculative at best. The reason that the use of a GC-Mass Spec is now standard procedure in fire investigation is that the risk of false positives (and, for that matter, false negatives) without one is so great, because it is so difficult to accurately determine a cause simply based on how a fully involved fire burned that even experts with decades of experience cannot do it properly.
Unsurprisingly, in the Willingham case, the state investigators never found evidence of an accelerant inside the house (a small sample consistent with charcoal lighter fluid was found on the door near where a bottle of fluid was stored, but none was found inside the house--this one positive sample is therefore not inconsistent with an accidental fire).
Todd Willingham's trial came during the period where fire investigation was being transitioned from an art to a science. In his case, the scientific method was simply not followed, and as a result false and misleading testimony was presented to the jury, leading to his conviction and execution. Another Texan, Ernest Willis, was released from death row a few months after Willingham was executed, after a review of his case found similar errors to the ones in Willingham's case (Willis was convicted in the mid-80s, and exonerated in 2004).
The Innocence Project estimates that there are between 100 and 200 people currently serving sentences for arson-related crimes who were convicted based on similar testimony to that which led to Willingham's execution. Inmates like Ed Graf, Daniel Dogherty (who now sits on death row in Pennsylvania awaiting execution) and Han Tak Lee have spent decades in prison based on this pseudo-science.
So what did [Todd Willingham's trial lawyer] David Martin's backyard experiment prove? First, it proved that someone intending to reproduce a burn pattern can do so by using an accelerant. It didn't prove by any stretch that the burn pattern in the Willingham home was caused by an accelerant, and it didn't prove that similar patterns couldn't be created in an accidental fire.
Second, and most importantly, it proved that Cameron Todd Willingham had a fool for a trial attorney, who neither understood the scientific method nor provided the sort of advocacy that all defendants are entitled to receive under the law.
Click here to read this feature in full.
Source:
Todd Willingham: On Burn Patterns and Stupid Lawyers, Daily Kos, Oct. 19, 2009
